Table 1.

Statistical tests.

Test useTestData structurePower
Psychophysical weighting calculated on all trials vs. only zero-mean trialsPearson correlationLinearSubjects: median r = 0.886 [0.819 to 0.952], 1 SEM; single session: median r = 0.846, [0.829 to 0.864], 1 SEM
Differences in slope of linear fit to temporal weights between flat-, late-, and early-stimulus conditions in humans and macaques (Fig. 3)Confidence intervalsLinearSlope of linear fit, [95% confidence interval]; Macaques: flat: –0.050, [–0.069 to 0.031]; late: 0.051, [0.004 to 0.098]; early: –0.094, [–0.111 to –0.077]; Humans: flat: –0.013, [–0.032 to 0.006]; late: 0.053, [0.006 to 0.100]; early: –0.083, [–0.119 to –0.048]
Comparison of slope of linear fit to temporal weights during early-stimulus condition between humans and macaque subjects (Fig. 3)Confidence intervalsLinearSlope of linear fit, [95% confidence interval]; Humans: –0.013, [–0.032 to 0.006]; Macaques: –0.050, [–0.069 to –0.031]
Comparison of psychometric functions across conditions (Fig. 3)Confidence intervalsLinearSlope of psychometric function, [95% confidence interval]; Macaques: early: 3.39 [3.22 to 3.56], flat: 2.16 [2.13 to 2.18], late: 2.9339 [2.83 to 3.03]; Humans: early: 2.77 [2.56 to 2.99], flat: 2.14 [2.00 to 2.28], late: 2.60 [2.43 to 2.77]
Average slope of temporal weights for flat-, early-, and late-stimulus conditions compared to 0 (Fig. 5)Wilcoxon sign testNon-Gaussianp < 0.0001, all conditions
Comparing group means for slopes of temporal weights for flat-, early-, and late-stimulus conditions (Fig. 5)ANOVAGaussianp < 0.0001
Comparing variance of slopes during flat stimulus condition vs. late- and early-stimuli (Fig. 5)Bartlett’s testNon-GaussianFlat-to-early, p < 0.0001; flat-to-late, p < 0.0001
Evaluating linear relationship between psychophysical threshold and slope of temporal weights (Fig. 5)Pearson correlationLinearFlat: r = –0.29, p < 0.001; early: r = 0.46 p = 0.038; late: r = 0.05, p = 0.75
Evaluating linear relationship between psychophysical threshold and energy of temporal weights (Fig. 5)Pearson correlationLinearFlat: r = 0.40, p < 0.0001; early: r = –0.004, p = 0.99; late: r = 0.31, p = 0.048